SAB-142 (fully human anti-thymocyte globulin)

The problem SAB-142 is trying to solve

Anti-thymocyte globulin (ATG) is one of the better-evidenced disease-modifying drugs in type 1 diabetes, and almost nobody talks about it. It is a decades-old transplant drug that depletes the T cells attacking the insulin-producing cells, and at a low dose — a tenth of the transplant dose — two randomized trials show that a single course preserves the body's own insulin production (measured as C-peptide) and lowers HbA1c for at least two years after diagnosis. Our full record on that drug is low-dose ATG.

The catch is in how it is made. Thymoglobulin is rabbit-derived: rabbits are immunized with human thymus cells, and the resulting rabbit antibodies are what get infused. A human immune system, predictably, notices. Serum sickness is common — 82% at the old 2.5 mg/kg dose, and still 32% at the lowest effective dose — and anti-drug antibodies build up. In practice that makes ATG a one-shot drug: you get a course, and a second course is neither comfortable nor obviously effective.

That matters, because an autoimmune attack is not a one-off event. A drug you can only give once can slow the decline. A drug you can give repeatedly might be able to hold it.

What SAB-142 is

SAB-142 is a fully human, multi-specific anti-thymocyte globulin from SAB Biotherapeutics. Instead of rabbits, it is produced in transchromosomic cattle — cattle engineered to carry human antibody genes, so the polyclonal antibodies they make are human proteins, not bovine or rabbit ones.¹

The mechanism is meant to be ATG's mechanism. In the phase 1, the effect tracked with exhaustion of CD4+ conventional T cells — the effector T cells that drive the attack — and the company describes the result as immune modulation without lasting depletion of the lymphocyte pool.¹ The intended difference from rabbit ATG is not potency. It is re-dosability: a human protein should not provoke the serum sickness and anti-drug antibodies that cap rabbit ATG at a single course.

That is a design rationale. Whether it holds in practice, over repeat courses, at scale, is exactly what the trials now running are for.

What the human data actually show — and their limits

The phase 1 (called HUMAN) was a randomized, double-blind, placebo-controlled, dose-escalation-and-redose study. In its type 1 diabetes cohort there were six participants: four given SAB-142 at 2.5 mg/kg and two given placebo, aged 19–40.¹

Read the next detail carefully, because it is the one most easily lost. Those participants had established T1D, diagnosed 28 to 40 months before they were randomized — not the newly diagnosed people the drug is being developed to treat.¹ What was reported:

• All four treated participants preserved C-peptide relative to baseline at day 120.
• Three of the four met a "super responder" definition — mean C-peptide at or above baseline through day 120.
• Mean CGM time in range rose from 73% to 85% by day 120, without an increase in insulin dose.¹

Those are genuinely striking numbers for people two to three years past diagnosis, when residual insulin production is usually falling, not rising. They are also four people. There is no control-arm comparison that can carry weight at that size, "super responder" is a label the company defined rather than a pre-registered endpoint, and every number above comes from press releases and conference presentations — the Immunology of Diabetes Society congress in April 2026 and an oral presentation at the ADA Scientific Sessions on 5 June 2026 — with no peer-reviewed publication.¹²³

So the honest position is: the phase 1 is a reason to run a phase 2b, not a reason to believe the drug works.

The trials that will settle it

SAFEGUARD (NCT07187531) is the one to watch. It is a randomized, double-blind, placebo-controlled phase 2b — the company calls it registrational, which is its own word rather than an FDA designation² — enrolling 159 participants at ~70 sites across 15 countries in the US, UK, EU, Australia and New Zealand. It has been recruiting since November 2025.⁴

Its design:

• Part A tests safety in ages 15–40; Part B tests efficacy in ages 5–40.
• Participants must be within 100 days of a stage 3 diagnosis — the honeymoon window, when there are still insulin-producing cells worth saving.
• Two infusion courses, six months apart — the re-dosing thesis, put to the test.
• The primary endpoint is C-peptide AUC from a 2-hour mixed-meal tolerance test at 12 months, the same currency teplizumab's trials trade in. Secondary endpoints include time in tight range, 70–140 mg/dL (3.9–7.8 mmol/L).⁴

Note what this means: SAB-142's only efficacy signal is in established T1D, and SAFEGUARD is testing it in new-onset T1D. Mechanistically that should favour the newly diagnosed, who have more to protect — but it is still an untested population for this drug.

The registry lists primary completion in November 2027; the company has guided to topline data in the second half of 2027.⁴ Either way, nothing decisive lands before late 2027.

Behind it, PRISE-hATG (NCT07670650) is a registered phase 3, not yet recruiting, planned to start in September 2026. It is sponsored by the University of Florida, with SAB Biotherapeutics and Breakthrough T1D as collaborators, and its scope is unusual: alongside new-onset participants carried over from SAFEGUARD, it enrolls recent-onset (100 days to 1 year) and established (1–2 years) T1D, ages 5–40, for a combined efficacy set of about 108.⁵ If the phase 1's established-disease signal is real, that trial is where it gets a proper test.

How to place it

Against teplizumab, which is FDA-approved for newly diagnosed children and has a phase 3 behind it, SAB-142 is far earlier and far less proven. Against rabbit ATG, which has two randomized trials behind it but no T1D approval and no realistic path to repeat dosing, SAB-142 is the same idea with the specific flaw engineered out.

The prize, if the re-dosing claim holds, is not a cure. It is turning beta-cell preservation from a one-time intervention into something maintainable — and, potentially, into something you could give earlier, before diagnosis. None of that is established. Today the whole case rests on four people and a design argument.

References

1. SAB BIO Presents Additional Clinical and Mechanistic Data from SAB-142 Phase 1 Trial in Adult Patients with Established Autoimmune Type 1 Diabetes at IDS 2026. SAB Biotherapeutics press release, 22 April 2026 (not peer-reviewed). https://www.globenewswire.com/news-release/2026/04/22/3278808/0/en/sab-bio-presents-additional-clinical-and-mechanistic-data-from-sab-142-phase-1-trial-in-adult-patients-with-established-autoimmune-type-1-diabetes-at-ids-2026.html ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶

2. SAB BIO to Present Data on SAB-142 at the American Diabetes Association's 2026 Scientific Sessions and FOCIS 2026 Annual Meeting. SAB Biotherapeutics press release, 29 May 2026. https://www.globenewswire.com/news-release/2026/05/29/3303527/0/en/sab-bio-to-present-data-on-sab-142-at-the-american-diabetes-association-s-2026-scientific-sessions-and-focis-2026-annual-meeting.html ↩ ↩²

3. ADA 2026 Recap: Days 1 and 2. Breakthrough T1D, June 2026. https://www.breakthrought1d.org/news-and-updates/ada-2026-recap-days-1-and-2/ ↩

4. SAFEGUARD — Phase 2b study of SAB-142 in stage 3 new-onset type 1 diabetes. NCT07187531, SAb Biotherapeutics. ClinicalTrials.gov (accessed 2026-07-14). https://clinicaltrials.gov/study/NCT07187531 ↩ ↩² ↩³

5. PRISE-hATG — Phase 3 study of SAB-142 in new-onset, recent-onset and established type 1 diabetes. NCT07670650, University of Florida with SAb Biotherapeutics and Breakthrough T1D. ClinicalTrials.gov (accessed 2026-07-14). https://clinicaltrials.gov/study/NCT07670650 ↩